Effer, a spare parts story.

Smaller update: tried several different ways to get the new-to-me compax3 drives to work with the PacSci servos that have resolvers. Between different encoders, phaking out the motor temp, and trying to get around the digital hall effect sensors the drives were not happy. After some machining jobs left the shop I decided to pony up and get some OG pac sci drives....and their cables!



If you are familiar with these drives... you will notice I have no option card in them, which is needed for setting the drives up. Fleabay to the rescue again. Actually found one on best-offer that was missing a hyphen in the part number so it was getting any looks! Struck a deal.

The connectors for the cables are military spec(Amphenol brand). My father-in-law, who worked on tanks in Germany for the US back in the day, immediately recognized them. I thought that was kinda cool.

Now to find a PC with Windows 97(3.1) to load the software....

 

Good find on the drivers and cables.
Ask your f-i-l if he worked on the M551 Sheridans. I did, they had many cannon plug connectors.

Edit: What controller did you get?
I found an instruction manual for a 400 series controller.

 

I didn't ask about models but showed in the plugs and he said he dealt with them every day!

I went with 900 series controllers. I was able to find manuals online but thank you!

 

If we are being honest. I mainly went with that AMB Spindle for the "quick change" tool-changing method and the replaceable brushes...and the 4 bolt-hole solid mount with dowels. You will really like being able to run bigger tools!

Static Rails vs Static Bearings
With static rails, you will have to compensate for reach with tooling. (or a base as you mentioned) Long-reach tooling has more flex than linear rails do. However, if you came up with something like a Pierson pallet system for a base this would be the way to go. My Fadal 4020 is like this. Big ole cast iron casting cantilevers the spindle center out some 212-14 inches to get half of the 20-inch travel.

I went the static bearing route so I could move up and clear things. I can also rotate my head 90 to clear....so maybe this is a moot point? I thought at the time that if I needed more stiffness I could move the bearings further apart or even add more bearings and lengthen the linear rails because I really don't care how tall it is. (Effer has to stand tall and proud to have a chance against his 5-ton shopmates!) You could set up your lowest bearings a few inches above the lowest part of your gantry to sneak in some room for tool height. My first VMC, Hitachi Seiki VM40H has static bearings and drops the "z plate" assembly (a massive casting). Whatever I have for tool-height I lose in travel though. Have snapped a few tools this way. lol

I would like to say make up an epoxy granite base with high side walls to perch your actual machine on (static bearing Z axis) but by the time you upgrade everything you might be looking at basically a new machine!

 

This was my intent. I know you could come up with a nice reliable low-buck option. Unless your foam machining needs to hold .0003 repeatability! Yes, this is the correct amount of zeroes.

I get this feeling. Maybe the best option is the one that will be easiest to manufacture and get you up and going? Keep you from sliding down the hill(slope) too much. lol

 

Been a minute since I measured but I think 33X22X10-20ish(depending on setup and 3/5 axis)

 

We will all be excited to see the progress. Speaking of...


The option card came in a handful of weeks ago. Seems to work as it should which is a relief. Installed Windows 95 on a virtual machine on my main laptop. It is a bit of a rain dance to get port forwarding to connect but I can talk to the option card. I BELIEVE I have one drive setup for step/direction correctly. As dumb as it sounds I haven't been able to test it out. Machining jobs have kept me at my other 2 machines. Though, currently working with a client on a project in which I will need Effer to be back up. Might be just the push we need!

 

You actually reminded me that I had an update. I hope I could shed some light on the subject for you.

 

So it looks like the Compax3 drives I picked up may not be the right match for PacSci servos, hence the PacSci drives. Not wanting the Parker brand Compax3 drives to go to waste, started looking at some Parker brand servos and cables from the Fleabay. Found some, though they are "customs." I would go into details but essentially after 2 afternoons of struggling ended up with this:



These work! HAve a better wiring/terminal solution on its way. Not near the punch of the PacSci servos but they are slightly stronger than all-in-one Chinese servos...annnd they are twice as fast! 6krpm constant vs 3k. SO, we can double up on reduction and get mo powa. Which is what we're after in the first place. These are also industrial servos/drives with nice features like autotune! Drives are smart and can handle their own homing procedures so could free up some inputs on the controller if necessary.

My end goal is still to use the PacSci setup. These will serve as an acceptable backup. Could also use these to run some 4th axis harmonic drive setups I am working on for my big CNCs too.

 

What we have here: The oldest pentium processor mother board combo running the oldest version of windows 2000nt. The pacsci servo software up and running. More importantly, talking directly to the drive over a direct (no usb buffers/coverters!) db9 cable. Most importantly, the drive is talking back! Able to upload/download and as you maybe able to see, reporting back encoder position in real time.

Feels like we kicked down one of the bigger doors holding up progress!

 

Been a couple months but made some progress last night!



This drive/servo fully powers up. I can setup the drives for velocith mode or step/direction and command them via the program/serial cable. Next step is too send signals from the Kflop controller. With success from that it will be time to start getting them mounted.

 

It's alive!

 

Error---wrong thread

 

Welded steel and Makita power tools?! You are the right thread!

 

Struggled a bit with real-world inputs to the drive. Both the Kflop controller and the 90's PacSci drives are smarter than I give them credit for!

I also needed to learn about internally biased differential inputs.

https://www.youtube.com/shorts/NtnESNYCwi4

Currently, they are running step/direction in a standalone fashion. Both the drives and controller can also run CCW/CW and quadrature. Which is crazy given the age difference and just the age of PacSci equipment.

Why would I need to run anything other than step/direction? Speed.

https://www.dynomotion.com/Help/StepAndDirection/StepAndDir.htm

This page walks through it nicely. In short: CCW/CW is faster than Step/Direction. Quadrature is the fastest of all three. So that's what I might do now that the wiring has been sorted.

Currently working on mounting the first servo to the Z axis. Then the other 2. All for the linear axis. I would imagine I will have to revisit my post processor once Effer is moving on all 5 again.

Worth mentioning; The drives have 2 inputs. 1 that stops only CW motion and the other that stops only CCW motion. Have thought about using these as hard limits but I really think I want this to be a soft stop or estop situation for the controller. Still cool though.

 

Another long overdue update:

I never thought this update would take years but uh, here we are. All 3 servos are mounted...crudely. Just weldments and bolt-ups.


(You can spy the thermal forming machine I built in the background)

Washers need to be placed to raise the bed to move over the bigger servo. Speaking of, just did some initial movements on the Z and it's quite snappy with the oversized motors/drives. The goal now is to get up and running. Either with 4 or 5 axis, to get some parts cast. I also have some prototype parts for clients to get moving on. Hence, the sudden wind of effort on Effer.

The 5-axis(right angle gearbox on y axis table) has issues. There is a slightly percetible slop that will just get magnified at length. I did snipe another harmonic drive but it needs a housing. However, just having the rotary on the spindle will allow cutting a dovetail, placing a sheet foam on its side, and window machine each side with a ball endmill in one setup. A win on its own.

The next step is to work on getting 220V power distribution setup and everything rewired. I will say, having found some OE cabling to go in between the drives and motors has greatly simplified things!